EDA - Clusters - 3BIO¶
In [160]:
# -*- coding: utf-8 -*-
In [161]:
%%HTML
<script src="require.js"></script>
In [162]:
# Importar librerÃas
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.pyplot import figure
import plotly.express as px
import seaborn as sns
import plotly.graph_objects as go
import plotly.io as pio
pio.renderers.default='notebook'
Data¶
In [163]:
from google.colab import drive
drive.mount('/content/drive/')
In [164]:
%cd /content/drive/My Drive/3BIO-Cluster-DataAnalysis
In [165]:
%ls
In [166]:
df = pd.read_csv("Cluster.cluster.csv")
df
Out[166]:
In [167]:
df.head(5)
Out[167]:
In [168]:
df.tail(5)
Out[168]:
In [169]:
df.describe()
Out[169]:
In [170]:
df.sort_values
Out[170]:
In [171]:
"""### Conocer estructura del DataFrame:"""
type(df)
Out[171]:
In [172]:
type(df["articulos"])
Out[172]:
In [173]:
df.columns
Out[173]:
In [174]:
# Nombres de Ãndices.
df.index
#list(df.index)
Out[174]:
In [175]:
# Imprimir los tipos de datos por columna (o por variable).
print(df.dtypes)
In [176]:
# Cantidad de columnas e imprimir su valor columnas
n_col=len(df.columns)
print("Cantidad de columnas: ",n_col )
In [177]:
# La Cantidad total de registros (filas)
n_reg= len(df.index)
print("Cantidad de registros : ", n_reg)
In [178]:
# Cantidad de cantidad de registros completos (sin los valores NA).
reg_com=len(df.dropna())
print("Registros completos sin NA", reg_com)
In [179]:
# Cantidad total de registros incompletos (filas con valores NaN).
reg_incom = n_reg - reg_com
print("Cantidad de registros incompletos: ", reg_incom)
In [180]:
# Clcular la cantidad de registros completos por variable.
nr_com= df.count()
print("Cantidad de registros completos por variable:\n ",nr_com)
In [181]:
# Cantidad de registros incompletos por variable.
print("Cantidad de registros incompletos por variable:")
print(n_reg - nr_com)
In [182]:
# Describir los valores estadÃsticos del DataFrame depurado.
df.describe()
df.plot()
df
Out[182]:
In [183]:
# Hacer estadÃsticas del cluster
df['Cluster'].describe()
Out[183]:
In [184]:
df.columns
Out[184]:
Generación de Nuevo Conocimiento¶
In [185]:
generacion = df.groupby('Cluster').sum().reset_index()[['Cluster','articulos','capitulos', 'innovaciones', 'libros', 'softwares']]
generacion[['articulos','capitulos', 'innovaciones', 'libros', 'softwares']]=generacion[['articulos','capitulos', 'innovaciones', 'libros', 'softwares']].div(generacion[['articulos','capitulos', 'innovaciones', 'libros', 'softwares']].sum(axis=1), axis=0).round(4) * 100
generacion
Out[185]:
In [186]:
fig = px.bar(generacion, x="Cluster", y =['articulos','capitulos', 'innovaciones', 'libros', 'softwares'], title="Generacion de nuevo conocimiento")
fig.show()
Instituciónes¶
In [187]:
instit = df.groupby('Cluster').sum().reset_index()[['Cluster','trabajos_grado', 'integrantes', 'doctorado', 'maestria','Especializacion', 'pregrado']]
instit[['trabajos_grado', 'integrantes', 'doctorado', 'maestria','Especializacion', 'pregrado']]=instit[['trabajos_grado', 'integrantes', 'doctorado', 'maestria','Especializacion', 'pregrado']].div(instit[['trabajos_grado', 'integrantes', 'doctorado', 'maestria','Especializacion', 'pregrado']].sum(axis=1), axis=0).round(4) * 100
instit
Out[187]:
In [188]:
fig = px.bar(instit, x="Cluster", y =['trabajos_grado', 'integrantes', 'doctorado', 'maestria','Especializacion', 'pregrado'], title="Clasificación Acádemica")
fig.show()
Clasificación de Investigadores¶
In [189]:
investigadores = df.groupby('Cluster').sum().reset_index()[['Cluster', 'Investigador_Asociado','Investigador_Emerito', 'Investigador_Junior', 'Investigador Senior']]
investigadores[['Investigador_Asociado','Investigador_Emerito', 'Investigador_Junior', 'Investigador Senior' ]]=investigadores[['Investigador_Asociado','Investigador_Emerito', 'Investigador_Junior', 'Investigador Senior',]].div(investigadores[['Investigador_Asociado','Investigador_Emerito', 'Investigador_Junior', 'Investigador Senior']].sum(axis=1), axis=0).round(4) * 100
investigadores
Out[189]:
In [190]:
fig = px.bar(investigadores, x="Cluster", y =['Investigador_Asociado','Investigador_Emerito', 'Investigador_Junior', 'Investigador Senior'], title="Clasificación de Investigadores")
fig.show()
In [191]:
sjr_cluster = df.groupby('Cluster').sum().reset_index()[['Cluster','SJR_Q_Q1','SJR_Q_Q2','SJR_Q_Q3','SJR_Q_Q4' ]]
sjr_cluster[['SJR_Q_Q1','SJR_Q_Q2','SJR_Q_Q3','SJR_Q_Q4' ]]=sjr_cluster[['SJR_Q_Q1','SJR_Q_Q2','SJR_Q_Q3','SJR_Q_Q4']].div(sjr_cluster[['SJR_Q_Q1','SJR_Q_Q2','SJR_Q_Q3','SJR_Q_Q4']].sum(axis=1), axis=0).round(4) * 100
sjr_cluster
Out[191]:
Scimago Ranking SJR Q¶
In [192]:
fig = px.bar(sjr_cluster, x="Cluster", y =['SJR_Q_Q1','SJR_Q_Q2','SJR_Q_Q3','SJR_Q_Q4'], title="Articulos Scimago Ranking SJR Q")
fig.show()
In [193]:
fig = px.bar(sjr_cluster, x="Cluster", y = ['SJR_Q_Q1'], title="Articulos Q1 en SJR",text_auto= True,labels={'x':'Clusters','y':'Articulos'},color_discrete_map={'SJR_Q_Q1':'blue'})
fig.show()
In [194]:
fig = px.bar(sjr_cluster, x="Cluster", y = ['SJR_Q_Q2'], title="Articulos Q2 en SJR",text_auto= True,labels={'x':'Clusters','y':'Articulos'},color_discrete_map={'SJR_Q_Q2':'red'})
fig.show()
In [195]:
fig = px.bar(sjr_cluster, x="Cluster", y = ['SJR_Q_Q3'], title="Articulos Q3 en SJR",text_auto= True,labels={'x':'Clusters','y':'Articulos'},color_discrete_map={'SJR_Q_Q3':'orange'})
fig.show()
In [196]:
fig = px.bar(sjr_cluster, x="Cluster", y = ['SJR_Q_Q4'], title="Articulos Q4 en SJR",text_auto= True,labels={'x':'Clusters','y':'Articulos'},color_discrete_map={'SJR_Q_Q4':'green'})
fig.show()
Participación de Clusters¶
In [197]:
fig = px.pie(df.groupby('Cluster').size().reset_index(), values=0, names='Cluster', title="Participación de Clusters")
fig.show()
Analisis de varianza¶
Cluster 0¶
In [198]:
df0 = df[df.Cluster == 0]
df0
Out[198]:
In [199]:
fig = px.box(df0, y=['articulos','capitulos','trabajos_grado'], points="all", color="Cluster")
fig.show()
In [200]:
fig = px.box(df0, y=['articulos','capitulos','trabajos_grado'], points="all", color='Instituciones')
fig.show()
In [201]:
fig = px.box(df0, y=['innovaciones', 'libros', 'softwares'], points="all", color="Cluster")
fig.show()
In [202]:
fig = px.box(df0, y=['innovaciones', 'libros', 'softwares'], points="all",color='Instituciones')
fig.show()
In [203]:
df0 = df0.sort_values(by=['articulos'], ascending=False)
df0h = df0.head()
df0h
Out[203]:
In [204]:
fig = px.pie(df0, values='articulos', names='Instituciones')
fig.show()
Cluster 1¶
In [205]:
df1 = df[df.Cluster == 1]
df1
Out[205]:
In [206]:
fig = px.box(df1, y=['articulos','capitulos','trabajos_grado'], points="all", color="Cluster")
fig.show()
In [207]:
fig = px.box(df1, y=['articulos','capitulos','trabajos_grado'], points="all", color='Instituciones')
fig.show()
In [208]:
fig = px.box(df1, y=['innovaciones', 'libros', 'softwares'], points="all", color="Cluster")
fig.show()
In [209]:
fig = px.box(df1, y=['innovaciones', 'libros', 'softwares'], points="all",color='Instituciones')
fig.show()
In [210]:
df1 = df1.sort_values(by=['articulos'], ascending=False)
df1s = df1.head()
In [211]:
fig = px.pie(df1s, values='articulos', names='Instituciones')
fig.show()
Cluster 2¶
In [212]:
df2 = df[df.Cluster == 2]
df2
Out[212]:
In [213]:
fig = px.box(df2, y=['articulos','capitulos','trabajos_grado'], points="all", color="Cluster")
fig.show()
In [214]:
fig = px.box(df2, y=['articulos','capitulos','trabajos_grado'], points="all", color='Instituciones')
fig.show()
In [215]:
fig = px.box(df2, y=['innovaciones', 'libros', 'softwares'], points="all", color="Cluster")
fig.show()
In [216]:
fig = px.box(df2, y=['innovaciones', 'libros', 'softwares'], points="all",color='Instituciones')
fig.show()
In [217]:
df2 = df2.sort_values(by=['articulos'], ascending=False)
df2.head()
Out[217]:
In [218]:
df2 = df2.sort_values(by=['articulos'], ascending=False)
df2s = df2.head()
fig = px.pie(df2s, values='articulos', names='Instituciones')
fig.show()
Cluster 3¶
In [219]:
df3 = df[df.Cluster == 3]
df3
Out[219]:
In [220]:
fig = px.box(df3, y=['articulos','capitulos','innovaciones','trabajos_grado'], points="all", color="Cluster")
fig.show()
In [221]:
fig = px.box(df3, y=['articulos','capitulos','innovaciones','trabajos_grado'], points="all", color='Instituciones')
fig.show()
In [222]:
fig = px.box(df3, y=['libros', 'softwares'], points="all", color="Cluster")
fig.show()
In [223]:
fig = px.box(df3, y=['libros', 'softwares'], points="all",color='Instituciones')
fig.show()
In [224]:
df3 = df3.sort_values(by=['articulos'], ascending=False)
df3.head()
Out[224]:
In [225]:
df3 = df3.sort_values(by=['articulos'], ascending=False)
df3s = df3.head()
fig = px.pie(df3s, values='articulos', names='Instituciones')
fig.show()
Cluster 4¶
In [226]:
df4 = df[df.Cluster == 4]
df4
Out[226]:
In [227]:
fig = px.box(df4, y=['articulos','capitulos','innovaciones','trabajos_grado'], points="all", color="Cluster")
fig.show()
In [228]:
fig = px.box(df4, y=['articulos','capitulos','innovaciones','trabajos_grado'], points="all", color='Instituciones')
fig.show()
In [229]:
fig = px.box(df4, y=['libros', 'softwares'], points="all", color="Cluster")
fig.show()
In [230]:
fig = px.box(df4, y=['libros', 'softwares'], points="all",color='Instituciones')
fig.show()
In [231]:
df4 = df4.sort_values(by=['articulos'], ascending=False)
df4.head()
Out[231]:
In [232]:
df4 = df4.sort_values(by=['articulos'], ascending=False)
df4s = df4.head()
fig = px.pie(df4s, values='articulos', names='Instituciones')
fig.show()
.
.
Borrador, INCOMPLETO HACIA ABAJO¶
.
Correlación¶
In [233]:
corr = df.corr()
mask = np.zeros_like(corr)
mask[np.triu_indices_from(mask)] = True
with sns.axes_style("white"):
f, ax = plt.subplots(figsize=(15, 15))
ax = sns.heatmap(corr,mask=mask,square=True,linewidths=.8,cmap="YlGnBu")
In [234]:
"""### Presentacion de los datos
"""
a = df['articulos']
Q1 = df['SJR_Q_Q1']
Q2 = df['SJR_Q_Q2']
Q3 = df['SJR_Q_Q3']
Q4 = df['SJR_Q_Q4']
In = df['Instituciones']
C = df['Cluster']
P = df['pregrado']
i = df['integrantes']
Ij = df['Investigador_Junior']
In [235]:
type(C)
Out[235]:
In [236]:
from IPython.core.interactiveshell import SpaceInInput
In [237]:
# Grafico de conjunto de datos de todas las variables
fig = plt.figure(figsize=(10,8), dpi=150)
ax1 = plt.subplot(3,3,1)
ax1.plot(C, Q1, lw='0.75', color='r')
ax1.set_title('Cluster vs Q1')
ax1.set_xlabel('Cluster')
ax1.set_ylabel('Q1')
ax2 = plt.subplot(3,3,2)
ax2.plot(C, Q2, lw='0.75', color='g')
ax2.set_title('Cluster vs Q2')
ax2.set_xlabel('Cluster')
ax2.set_ylabel('Q2')
ax3= plt.subplot(3,3,3)
ax3.plot(C, Q3, lw='0.75', color='b')
ax3.set_title('Cluster vs Q3')
ax3.set_xlabel('Cluster')
ax3.set_ylabel('Q3')
ax4 = plt.subplot(3,3,4)
ax4.plot(C, Q4, lw='0.75', color='c')
ax4.set_title('Cluster vs Q4')
ax4.set_xlabel('Cluster')
ax4.set_ylabel('Q4')
#ax5 = plt.subplot(3,3,5)
#ax5.plot(a, In, lw='0.75', color='m')
#ax5.set_title('Instituciones con mas articulos')
#ax5.set_xlabel('Articulos')
#ax5.set_ylabel('Instituciones')
# ax6 = plt.subplot(3,3,6)
# ax6.plot(i, Ij, lw='1.50', color=[0.2,0.25,1.0])
# ax6.set_title('Investigadores Jr entre integrantes')
# ax6.set_xlabel('Integrantes')
# ax6.set_ylabel('Investigadores Jr')
# Prevenir la superposicion de los titulos de subplots con los ejes contiguos
plt.tight_layout()
fig.subplots_adjust(top=0.90)
plt.show()
In [238]:
#%%shell
#jupyter nbconvert --to html /EDA_Clusters_3BIO.ipynb